AI Engineer Salary: What the Pay Data Shows and What a Degree Costs

The Bureau of Labor Statistics has no wage line called AI engineer, so there is no official salary under that title. It counts AI engineers with software developers, so these figures cover that wider group. The May 2025 median is $135,980 a year, and the middle half earns $105,210 to $171,980. Your state moves the number more than almost anything else.

Why Is There No Official AI Engineer Salary?

The federal wage survey sorts every job into a fixed list of occupation codes, and that list is revised only every few years. Employers coined “AI engineer” faster than the list could follow, so BLS files the work under software developers (code 15-1252).

Keep that in mind when a job site quotes one tidy AI engineer figure. Numbers like that tend to come from postings or user-submitted salaries, which skew toward the high end and can’t be compared across years. Here we rely on the federal survey alone.

The survey covers 1,687,890 software developers nationwide. Some of them build AI products and most do not, and the survey can’t tell them apart. So use the numbers below as the range your job sits in, not as a quote for one title.

If you are still deciding whether the job fits you, our AI engineer career page explains what the job involves and which degree level employers expect.

What Is the Median AI Engineer Salary?

For software developers, the May 2025 national median is $135,980 a year, which works out to $65.38 an hour. The mean is $148,100, or $71.20 an hour.

The mean runs higher than the median because well-paid workers at the top stretch the average. For planning, lean on the median: half of these workers are paid more than it, and half are paid less.

Here is how the pay spreads out across the whole group.

• 10th percentile: $82,460.
• 25th percentile: $105,210.
• Median: $135,980.
• 75th percentile: $171,980.
• 90th percentile: $214,670.

A quarter of software developers earn $105,210 or less, and a quarter earn $171,980 or more. That wide middle band shows how much your own situation matters, and it is the reason a single salary figure for the whole job can mislead you.

How Does Experience Change Pay?

BLS does not publish wages by years of experience, and we won’t invent year bands. The percentiles are the honest stand-in, because experience, skill and employer all feed into where a worker lands.

Read them as a ladder. The 10th percentile of $82,460 covers people early in their careers and people in lower-paying markets. The 75th percentile of $171,980 and the 90th of $214,670 take in senior engineers, people in high-pay states and people at employers that pay well.

No survey can tell you which rung you will start on. Look at postings in your city for the entry level, then compare them to this ladder.

Which States Pay AI Engineers the Most?

Where you work matters as much as what you know. These are the five states with the highest annual median wage for software developers, with the number of workers there.

• California: median $174,410, middle half $136,990 to $216,670, with 284,390 workers.
• Washington: median $166,540, middle half $132,290 to $211,860, with 107,030 workers.
• New York: median $166,180, middle half $127,160 to $205,540, with 113,510 workers.
• Massachusetts: median $165,210, middle half $131,370 to $190,280, with 48,190 workers.
• Oregon: median $142,720, middle half $111,750 to $170,600, with 21,830 workers.

Maryland follows with a median of $138,680, then Colorado at $138,390 and Virginia at $136,460. Texas has the second-most software developers, 163,880, with a median of $132,150. Florida has 87,890 and a median of $130,980.

Ranked by mean wage, the top states are California at $186,770, Washington at $171,040, New York at $163,820, Massachusetts at $161,730 and Maryland at $151,560.

California’s 25th percentile of $136,990 is higher than the national median. That is how large the location effect can be.

Which States Pay the Least?

The lowest medians are in Mississippi at $95,330, South Dakota at $96,310, Louisiana at $102,800 and Arkansas at $104,490. Indiana’s is $105,920 and North Dakota’s is $105,600.

These states also employ far fewer developers. Mississippi has 1,880 and South Dakota has 2,380, against California’s 284,390.

Lower pay is not the whole story. Rent, taxes and the price of everyday goods run lower in many of these places, so a smaller paycheck can go further than the numbers suggest. The survey does not adjust for that, and neither do we.

Does Remote Work Change the Picture?

It can. Many online AI degree holders look for jobs they can do from home, and pay policies vary. Some employers set pay by the market where the company is based, and others set it by where you live.

If you live in a lower-paying state and get hired by a company based in a higher-paying one, your pay may land somewhere in between. The survey measures where the job is located, not where an employer’s headquarters sit, so we can’t say how often that happens. Ask how pay is set before you accept an offer, and whether it changes if you move later.

What About the Research End of AI Work?

Some AI jobs are research-style roles, where you design and test new methods. BLS counts those workers separately, as computer and information research scientists.

That group is much smaller, at 37,200 workers in May 2025. The median is $140,300, the 25th percentile is $103,570, the 75th is $188,700 and the 90th is $230,630. BLS lists a master’s degree as the typical entry education.

So the two ends of AI work sit close together at the median. The research group’s top end runs higher, and its entry bar is higher too. Our machine learning engineer salary page covers a third nearby group, data scientists.

How Fast Is Demand Growing?

BLS projects employment of software developers to grow 10 percent from 2025 to 2035, which it calls much faster than average. That adds about 185,400 jobs to a base of 1,905,400. BLS also expects about 106,100 openings a year, which include replacements for people who leave.

Research scientists are projected to grow 22 percent, from a much smaller base of 38,600 jobs.

Projections are estimates. They say nothing about what any one employer will pay, and BLS revises them each cycle. Growth also doesn’t mean easy hiring, because entry-level roles draw many applicants.

What Does an Online AI Degree Cost Next to This Pay?

The right comparison is the program’s total against the pay range for the job you are after, not against a salary you saw in an ad. Online AI master’s programs on our list have a median total tuition of $32,463. Bachelor’s programs have a median of $58,800.

These examples span the price range, using each school’s published total before fees.

• The University of Texas at Austin: $9,990 for 30 credits, with no GRE required. See the online MS in Artificial Intelligence.
• University of Colorado Boulder: $15,750 for 30 credits.
• West Virginia University: $24,150 for 30 credits at $805 per credit.
• Saint Louis University: $24,300 for its applied AI master’s.
• Purdue University: $28,500 for 30 credits at $950 per credit.
• Penn State World Campus: $32,670 for its AI Engineering option.
• University of Pennsylvania: $38,590 for the MSE in AI.
• Duke University: $76,650 for 30 credits, with some onsite time.
• Columbia University: $81,000 for 30 credits at $2,700 per credit.

From cheapest to costliest, the price differs by roughly eight times. Pay in the survey isn’t broken out by school, so nothing here shows that a pricier program earns a bigger paycheck.

At the bachelor’s level, Dakota State charges $373.80 per credit, which totals $44,856 for 120 credits. Our list of the cheapest online AI degrees sorts every level by price, and the online master’s ranking shows format and GRE rules next to cost.

How Do You Weigh the Cost Against Your Own Pay?

The same degree can be a small expense or a big one, depending on where you start. A software developer who already earns the national median has a short gap to close. Someone moving in from a field that pays less has a bigger one.

Do the math with your own numbers. Take the program’s total, add fees and any time away from work, then look at the pay range for your target job in your state.

The lowest-priced master’s on our list costs less than a tenth of the national median software developer wage. The highest-priced one costs more than half of it. That difference can change the decision on its own.

Our AI degree value guide takes you through that comparison. If you want the engineering path, the online AI engineering degrees ranking covers programs built around it.

How Can You Raise Your Pay in AI Work?

The data supports three levers, and only three.

Location. The state spread is the largest effect in the survey. California’s median is $174,410 and Mississippi’s is $95,330. If you can move, or work for an employer in a higher-paying state, that can matter more than any one credential.

Type of work. Research-style roles sit under a different occupation, with a higher entry bar and a higher 90th percentile of $230,630.

Education. BLS ties a master’s to research scientist entry and a bachelor’s to software developers. A degree can open doors, but the survey does not report pay by degree, so we can’t promise a raise.

What Does the Hourly Rate Look Like?

Coming from a job paid by the hour? Hourly pay is the easier match for your current wage, and BLS reports it in the same survey. For software developers, the national median is $65.38 an hour and the mean is $71.20.

Two states show how far the rate can swing. California’s median hourly rate is $83.85, and Mississippi’s is $45.83. A full-time schedule is about 2,080 hours a year, so a few dollars an hour adds up quickly over a year of work.

Hourly figures also help if you plan to freelance. Self-employed workers pay their own taxes, benefits and slow weeks, so a contract rate has to cover more than a salaried hour does. The survey covers wage earners, so it won’t tell you what a contractor should charge.

What Are the Limits of This Data?

Keep four caveats in mind when you use these figures.

BLS tallies occupations and ignores job titles. One company may file an AI engineer under software developers while another uses a different code.

Every worker in the group goes into the same pool. A developer with two decades on the job and a recent graduate share one median.

Only wages are measured, not whole compensation packages. Bonuses and stock grants are absent from our data, so we leave them out.

Finally, a few states have wage figures suppressed when BLS can’t publish them reliably. Alaska, for example, has an employment count but no published median for software developers.

The figures still give you a solid frame. Build a range from them, then test it against current postings where you plan to work.

What Should You Do With These Numbers?

Read them as a range. Find your state, note its percentiles, and hold them up against the price of the program you are weighing.

Next, find people who do this work and ask a few plain questions: what title they were hired under, what they studied, and how their pay lines up with the survey. One employer’s pay scale is beyond what a national dataset can show.

If you want the job description first, read our AI engineer career guide. If you are still working out the path, see how to become an AI engineer.